Is ventilated hospital-acquired pneumonia a worse entity than ventilator-associated pneumonia?

Epidemiology

The incidence of VAP varies according to the characteristics of the study population [8], diagnostic criteria [9] and geographical origin. Dudeck et al. [10] reported an incidence of VAP rate of 1–2.5 cases per 1000 ventilator-days in North America. Conversely, Koulenti et al. [11] observed an incidence of 18.3 VAP episodes per 1000 ventilator-days in Europe. A recent article reported the incidence of HAP as 3.63 cases per 1000 patient-days in the United States [12]. In a prospective study on ICU-acquired pneumonia [13], 52% of cases were diagnosed as VAP and 48% as HAP. Specifically, >50% of HAP cases required mechanical ventilation. Mortality and outcomes vary significantly between the different subtypes of nosocomial pneumonia [14]. In 2016, Micek et al. [15] observed that patients with HAP presented an 8.4-fold increased risk of death, an eight-fold increase in mechanical ventilation and a longer hospital length of stay compared with those not affected by HAP. More recently, seven clinical trial datasets on both VAP and HAP in patients admitted to ICU have been analysed [16]. V-HAP had the highest all-cause mortality rate at day 28 (27.8%), followed by VAP (18%) and NV-HAP (14.5%).

Although the reasons for this difference in mortality are still unknown, and further studies are warranted to shed light on this matter, we believe that several factors may be involved. First, the spread of noninvasive ventilation (NIV) and high-flow oxygen therapy (HFOT) reduces the percentage of patients requiring immediate intubation, but increases the number of sick patients with risk factors for acquiring severe HAP [17] who could require subsequent intubation and mechanical ventilation. Moreover, compared with VAP, HAP poses additional challenges for respiratory sampling and, thus, for obtaining microbiological diagnosis [18], possibly causing a delay in the appropriate treatment, especially on the ward. Indeed, it has been demonstrated that rapid diagnosis and initiation of adequate empiric antimicrobial treatment are of paramount importance in the outcome of patients with nosocomial pneumonia [19–21]. Lastly, many care packages to prevent VAP have been implemented recently, leading to a reduction in incidence of VAP [1, 2, 17]. Their efficacy in improving the survival of patients with VAP has been demonstrated [22]. Conversely, little work has been done on HAP prevention, and the efficacy of most potential strategies remain unproven [23].

Pathophysiology

The pathophysiology of nosocomial pneumonia relies on the paradigm of pharyngeal bacterial aspirations, which considers the healthy distal airways to be sterile site that is contaminated by micro-aspiration of digestive contents during the hospital stay. Furthermore, patients admitted to hospital demonstrate a severe impairment of respiratory physiology [24], which is influenced by the patient's medical conditions (e.g. main disease and comorbidities) and therapeutic interventions (e.g. drugs or mechanical ventilation) [25]. The microbiological evidence of nosocomial pneumonia was more frequently available in patients with VAP compared with HAP [18]. However, in patients for whom the microbiological aetiology was known, pathogens were similar between VAP and HAP, suggesting that both types of patients might benefit from similar antimicrobial treatment.

Risk factors for multidrug-resistant pathogens

The dramatic increase of multidrug-resistant (MDR), extensively drug-resistant and pandrug-resistant bacteria [26] has complicated the management and outcome of nosocomial pneumonia [27]. Empiric treatment of VAP and HAP should be guided by local epidemiology and susceptibility data, together with a careful assessment of individual risk factors for MDR pathogens. The 2016 Infectious Diseases Society of America (IDSA)/American Thoracic Society (ATS) guidelines [1] and the 2017 European Respiratory Society (ERS)/European Society of Intensive Care Medicine (ESICM)/European Society of Clinical Microbiology and Infectious Diseases (ESCMID)/Latin American Thoracic Association (ALAT) guidelines [2] provided a different algorithm for the empiric therapy of nosocomial pneumonia, based on patient- and centre-oriented risk factors for MDR pathogen infection [28]. According to the 2016 IDSA/ATS guidelines [1], risk factors for MDR infection in patients with VAP included previous intravenous antibiotic therapy within 90 days, hospitalisation for ≥5 days before the diagnosis of VAP and organ failure (septic shock at the time of VAP diagnosis, acute respiratory distress syndrome and need for renal replacement therapy before VAP onset). Risk factors for MDR infection in patients with HAP have been investigated less, and only previous use of i.v. antibiotics has been identified in patients with HAP, without any distinction between NV-HAP and V-HAP [29]. This risk factor has been identified as the most predictive for VAP/HAP, specifically caused by methicillin-resistant Staphylococcus aureus (MRSA) [30, 31] and Pseudomonas aeruginosa [32].

In a validation study of the 2016 IDSA/ATS guidelines including 316 patients with VAP and HAP, Ekren et al. [33] observed a higher sensitivity compared with specificity for the majority of the risk factors reported. Previous i.v. antibiotic use and ≥5 days of hospitalisation showed the highest negative predictive values for VAP (79% and 80%, respectively) and i.v. antibiotic use for HAP (69%). Based on the role of local ecology [34] and the growing evidence that colonisation with MDR pathogens influences the occurrence of MDR pneumonia [35, 36], the 2017 ERS/ESICM/ESCMID/ALAT guidelines suggested taking into account hospital settings with high rates of MDR pathogens (>25%) and previous colonisation with MDR pathogens to identify patients at increased risk of MDR infection [37]. However, the authors did not include acute respiratory distress syndrome and renal replacement therapy in the definition, since they have been related to severity of the disease rather than the risk for MDR pathogen infection. Nevertheless, we consider the risk of mortality due to the degree of organ dysfunction in order to implement empiric antibiotic therapy [38].

Diagnosis

The diagnosis of nosocomial pneumonia is based on the presence of a new pulmonary infiltrate on chest radiography or computed tomography together with some clinical features (such as fever or hypothermia, purulent secretions, oxygenation impairment) and laboratory data (leucocytosis or leukopenia). Theoretically, antibiotic choice should be based on the results of cultures of respiratory secretions [1]. Of course, if empiric antimicrobial therapy is necessary, the collection of a lower respiratory tract sample should be undertaken prior to the beginning of empiric treatment.

As discussed earlier, a higher rate of mortality in patients with bacterial V-HAP than those with bacterial VAP has been reported recently [16]. Indeed, HAP still represents a challenge for clinicians due to the difficulty of collecting a lower respiratory tract sample [7]. This can be done using an invasive technique such as fibreoptic bronchoscopy to collect a bronchoalveolar lavage. However, it requires the expertise of qualified clinicians and is not risk-free (e.g. pneumothorax, impact on gas exchange, haemodynamic instability). Due to these potential deleterious complications, the benefits of this technique are not clear [1]. Moreover, while it is easy and feasible to access the lower airways through an endotracheal tube, performing a bronchoscopy in a spontaneously breathing patient can be challenging, although several reports have proved the safety of this technique in patients on NIV or HFOT [39–42].

Alternatively, the culture of sputum is a noninvasive technique that can be obtained spontaneously, by induction via nasotracheal suctioning or endotracheal aspiration in intubated patients. Data on this practice are limited [7], due to the difficulty of obtaining an adequate sample and the need for patient collaboration and an assessment of quality and representativeness of the lower respiratory tract. Moreover, a positive result may only represent colonisation [7].

No recommendation was made on invasive or noninvasive diagnostic methods in VAP in the 2016 IDSA/ATS guidelines [1]; since their impact on the outcome is comparable, there is no evidence regarding sampling in HAP, even though the same guidelines underline the importance of noninvasive respiratory sampling in nosocomial pneumonia rather than empiric treatment alone. This recommendation is graded as weak with low-quality evidence, based on a lack of research showing an improvement in patient-centred outcomes. Nevertheless, culture data may represent an opportunity for increasing appropriate coverage, thus minimising unnecessary antibiotic exposure. Accordingly, the collection of sputum must be encouraged in patients with nonventilated HAP, ensuring the quality and lower respiratory tract representativeness. This recommendation mainly comes from evidence on community-acquired pneumonia (CAP) [43], although some reports from HAP are promising [44].

In a recent observational study including 200 patients with HAP, 122 of whom required mechanical ventilation, patients assessed with an invasive diagnostic method had higher proportions of microbiological diagnosis [18]. Patients with microbiological diagnosis more frequently presented changes in their empiric antibiotic scheme, mainly de-escalation. Interestingly, the authors observed that severity, hypoxaemia and chest radiography patterns were associated with intubation after performing a fibreoptic bronchoscopy; in addition, they suggested the need to develop a prediction tool to evaluate which patients with HAP would need mechanical ventilation.

It has been demonstrated that almost all the severity scores used for CAP could be also valuable for HAP [45]. A more recent study focused only on the CURB-65 score, named for the risk factors measured (each scoring 1 point): confusion of new onset, urea ≥7 mmol·L⁻¹ (blood urea nitrogen >19 mg·dL⁻¹), respiratory rate ≥30 breaths per minute, systolic blood pressure (BP) ≤90 mmHg or diastolic BP ≤60 mmHg and age ≥65 years [46]. A score ≥2 could help to stratify patients with an increased risk of mortality [47]. The use of this severity score in HAP needs further validation, but it could be an important tool to improve the delivery of care for pneumonia patients and to identify early those patients who could deteriorate and need mechanical ventilation [16].

Aetiology and treatment

The development of HAP and VAP is due to an imbalance between normal host defence and the ability of micro-organisms to colonise and then invade the lower respiratory tract, which is increased in the presence of new points of entrance for the bacteria, such as central venous catheters, nasogastric tubes and tracheostomies. In addition, patients with VAP are usually exposed to more antibiotics than those with HAP, either for past infection or prophylaxis, causing increased colonisation with nosocomial pathogens. Indeed, it has been reported that changes in the oropharyngeal flora in patients with HAP may be delayed, with a persistence of “community-acquired microbes” such as Streptococcus pneumoniae, while patients with VAP have a higher prevalence of nonfermenting and enteric Gram-negative bacilli and methicillin-sensitive S. aureus (MSSA) [13, 48, 49]. However, a recent small retrospective study by Feng et al. [50] of the microbial aetiology of HAP and VAP reported that the leading pathogens were similar in both entities, with a predominance of Acinetobacter baumannii and P. aeruginosa in VAP, and Klebsiella pneumoniae and S. aureus in HAP. Moreover, the authors demonstrated an increase in the MDR bacteria in VAP, with a lower susceptibility of P. aeruginosa in VAP compared to HAP, as reported in previous studies [51, 52]. The results of Feng et al. [50] are in line with those of a less recent prospective observational study by Esperatti et al. [13] about ICU-acquired nosocomial pneumonia, where the proportion of relevant pathogens in patients with an isolated organism was similar, suggesting that the cause and the outcome depend on the underlying conditions of the patients rather than previous intubation. To corroborate this hypothesis, this group published another study showing no differences in terms of systemic inflammatory response, severity of nosocomial pneumonia and clinical characteristics between MDR and non-MDR cases. Of note, these findings were similar for VAP and NV-HAP [53]. Therefore, it seems that aetiology and outcomes are similar regardless of whether nosocomial pneumonia is acquired or not during ventilation [13, 54], and the appropriate empiric treatment should take into account clinical risk factors for MDR germs, other than the severity of the patient. Consequently, the 2017 European guidelines [2] did not provide a strict distinction in terms of risk factors between VAP and HAP, and it is considered a good practice to apply the recommendations used for VAP in patients with V-HAP and NV-HAP. Treatment should only be based on the risk of mortality and the likelihood of MDR pathogens [2] (table 1). If the risk of MDR is low with a mortality risk <15%, monotherapy with a narrow-spectrum antibiotic against nonresistant Gram-negative bacteria and MSSA would be an effective choice. However, in patients with a high risk of mortality, a single Gram-negative agent active against Pseudomonas would be an effective choice (except aztreonam, which does not cover Gram-positive bacteria) if the patient is not in septic shock. Conversely, patients with septic shock should always be treated with two antipseudomonal antibiotics, including an antipseudomonal β-lactam (imipenem, meropenem, cefepime, piperacillin/tazobactam, ceftazidime and aztreonam) plus a second agent such as an aminoglycoside (preferably amikacin, which is the most active agent in many ICUs) or an antipseudomonal quinolone (ciprofloxacin or levofloxacin). Note that in ICUs with high rates of MRSA, linezolid or vancomycin should be added. Similarly, when the prevalence of Acinetobacter spp. and extended-spectrum β-lactamases (ESBL)-producing Enterobacteriaceae is high, the regimen should include also dual anti-Pseudomonas cover. If Acinetobacter spp. are commonly isolated, the second agent should be colistin. For ESBL-producing germs, a carbapenem may be the best choice, and cefepime and piperacillin/tazobactam only considered based on local susceptibilities. The dual coverage should not be continued after the results of the cultures, except in patients with MDR pathogens. The duration of treatment should be based on the clinical response, the presence of an inappropriate initial empiric therapy or MDR organisms and serial measurement of biomarkers. Usually, 7–8 days are effective in patients without immunodeficiency, cystic fibrosis, empyema, lung abscess, cavitation or necrotising pneumonia and with a good clinical response to therapy.